-
Notifications
You must be signed in to change notification settings - Fork 86
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
about trainning by Y channel with yuv files #17
Comments
Sorry, could you please provide more details about your question? Do you mean you want to downscale and upscale images with only Y channels? And what kind of advice do you want? |
sorry ,yes i want to downscale and upscale images(just Y channel) with only Y channels ,and i have tried to train the model with Y channel of DIV2K YUV files, I changed the "in_nc" and "out_nc" to '1',then changed the index to '1' , like the "self.output[:, :3, :, :]"to "self.output[:, :1, :, :]", i think when the channel is 3 ,the output of netG is [:,12,:,:],and change to be [:,4::,:], is it right ? now i have trained about 100000, but the result is very blur , maybe Something important have been missed , can you give me some |
sorry ,yes i want to downscale and upscale images(just Y channel) with only Y channels ,and i have tried to train the model with Y channel of DIV2K YUV files, I changed the "in_nc" and "out_nc" to '1',then changed the index to '1' , like the "self.output[:, :3, :, :]"to "self.output[:, :1, :, :]", i think when the channel is 3 ,the output of netG is [:,12,:,:],and change to be [:,4::,:], is it right ? now i have trained about 100000, but the result is very blur , maybe Something important have been missed , can you give me some advice? |
If you have correctly modified the dataloader I think your code is right. About the result, how blur is it? (e.g. how much PSNR?) By the way, the pretrained models can also be used for grayscale images, for example, there exists a grayscale image in Set14 I remember, and if we read the image by opencv which changes the image into three channels (just copying the value of gray channel), results are good as well. You may test it and consider copying Y channel into three channels. |
thank you so much , i will check my dataloader and test the grayscale image , and copy Y to three channels , I ignored the PSNR later , i will check it too. i will share all the results to you after the next trainning, thanks for your guidance ! |
hi amazing work
I want to train use Y channel with yuv files , can you give me some advices about it ?
The text was updated successfully, but these errors were encountered: